In situ field measurements as well as targeted laboratory studies have shown that freeze-thaw cycles (FTCs) affect soil trace gas fluxes. However, most of past laboratory studies adjusted soil moisture before soil freezing, thereby neglecting that snow cover or water from melting snow may modify effects of FTCs on soil trace gas fluxes. In the present laboratory study with a typical semi-arid grassland soil, three different soil moisture levels (32 %, 41 %, and 50 % WFPS) were established (a) prior to soil freezing or (b) by adding fresh snow to the soil surface after freezing to simulate field conditions and the effect of the melting snow on CO 2 , CH 4 , and N 2 O fluxes during FTCs more realistically. Our results showed that adjusting soil moisture by watering before soil freezing resulted in significantly different cumulative fluxes of CH 4 , CO 2 , and N 2 O throughout three FTCs as compared to the snow cover treatment, especially at a relatively high soil moisture level of 50 % WFPS. An increase of N 2 O emissions was observed during thawing for both treatments. However, in the watering treatment, this increase was highest in the first thawing cycle and decreased in successive cycles, while in the snow cover treatment, a repetition of the FTCs resulted in a further increase of N 2 O emissions. These differences might be partly due to the different soil water dynamics during FTCs in the two treatments. CO 2 emissions were a function of soil moisture, with emissions being largest at 50 % WFPS and smallest at 32 % WFPS. The largest N 2 O emissions were observed at WFPS values around 50 %, whereas there were only small or negligible N 2 O emissions from soil with relatively low soil water content, which indicates that a threshold value of soil moisture might exist that triggers N 2 O peaks during thawing.